The intention is to develop an interactive model of how the immune system operates to deal with a respiratory infection, and to use this to test ideas about the nature of cell-mediated immunity (CMI). The initial aim is to make a rigorous analysis of the cellular changes in the regional lymph node and lung inflammatory exudate of mice with influenza. This experimental system will then provide the basis for assessing the consequences of eliminating particular lymphocyte subsets, cytokines or segments of the T cell repertoire. The parameters that will be examined include modification of lymphocyte recruitment patterns, the generation of specific effector functions, regulation of the clonal expansion and differentiation of T cells and the involvement of non-specific elements of the host response. The need is to understand how these various cell populations fit together to promote effective immunity. The technology that is required to make such analyses is now available. Preliminary studies of influenza in the mouse demonstrate assessment of lymphocyte specificity and effector capacity can be used to develop a quantitative understanding of the total CMI response in the mediastinal lymph node and the lung. The emphasis taken here depend on the fact that recent advances in biotechnology, such as the capacity to produce large amounts of recombinant cytokines, and neutralizing monoclonal antibodies, offer immense potential for manipulating the immune response. However, the application of therapeutic protocols that involve the inoculation of massive doses of lymphokines has indicated that the consequences of such treatments are both complex and poorly understood. It is also apparent that many of the postulates that have been developed from in vitro studies in lymphocyte culture systems cannot be translated in a naive way to the situation in the intact host. There is obviously a major need for the development of model systems that look at the host response in vivo in a comprehensive and rigorous way. The combined application of contemporary phenotypic and functional analysis used here for influenza has not been attempted previously for any respiratory infection, and is already giving a unique picture of the interactive events that determine CMI. The experiments have broad implications for understanding the host response to viruses.